On Mon, Aug 05, 2013 at 11:52:20AM +0300, Martin Storsjö wrote:
> On Mon, 5 Aug 2013, Diogo Franco (Kovensky) wrote:
> >On 8/4/2013 2:23 PM, Martin Storsjö wrote:
> >>Diogo, does this part actually fix an issue itself, or is it
> >>just to avoid needless workarounds? Does it harm anything if
> >>this is kept as is, to simplify the conditions? (That is, since
> >>cygwin does support the win32 api, the function is available and
> >>we'd hope that using it should work as intended on cygwin as
> >>well.)
> >
> >Command line argument parsing is broken at least on cygwin64
> >without the patch. I'm not sure *why* it happens though; I thought
> >it'd be a winapi misuse (e.g. using C types instead of winapi
> >defines; win64 has 32bit long but cygwin64 has 64bit long) but it
> >didn't seem to be the case (no gcc warnings at least), however
> >simply commenting out the call to prepare_app_arguments in
> >cmdutils.c magically made it work, so the cause must be the
> >windows-specific command line argument parsing.
> 
> I had a closer look, and it seems that GetCommandLineW() only gives
> the exe name but none of the parameters. And this only happens on
> cygwin64, on cygwin32 it seems to work just fine.
> 
> While we could work around it, the question is whether this is an
> intentional decision from cygwin's side to break this function or if
> it's a bug they might be able to fix?

Before we start littering our code with workarounds for such issues we
should indeed rule out that it's a bug.  If it is a bug, the bug should
be reported and hopefully get fixed, obviating the need for workarounds.

Diego
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to